17th National Conference of the Australian Meteorological and Oceanographic Society
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 11 (2010) 012031
doi:10.1088/1755-1315/11/1/012031
Teleconnections associated with the intensification of the
Australian monsoon during El Niño Modoki events
A S Taschetto1, R J Haarsma2, A Sen Gupta1, C C Ummenhofer1 and M H
England1
1
Climate Change Research Centre, University of New South Wales, Sydney, NSW
2052 Australia
2
Royal Netherlands Meteorological Institute, De Bilt, The Netherlands
E-mail: [email protected]
Abstract. In this study we investigate the teleconnection between the central-western Pacific
sea surface temperature (SST) warming, characteristic of El Niño Modoki events, and
Australian rainfall using observations and atmospheric general circulation model experiments.
During Modoki events, wet conditions are generally observed over northwestern Australia at
the peak of the monsoon season (i.e. January and February) while dry conditions occur in the
shoulder-months (i.e. December and March). This results in a shorter but more intense
monsoon season over northwestern Australia relative to the climatology. We show that, apart
from the well-known displacement of the Walker circulation, the anomalous warming in the
central-western equatorial Pacific also induces a westward-propagating disturbance associated
with a Gill-type mechanism. This in turn generates an anomalous cyclonic circulation over
northwestern Australia that reinforces the climatological mean conditions during the peak of
the monsoon season. The anomalous circulation leads to convergence of moisture and
increased precipitation over northern Australia. This response, however, only occurs
persistently during austral summer when the South Pacific Convergence Zone is
climatologically strengthened, phase-locking the Gill-type response to the seasonal cycle. The
interaction between the interannual SST variability during El Niño Modoki events and the
evolution of the seasonal cycle intensifies deep convection in the central-west Pacific, driving a
Gill-type response to diabatic heating. The intensified monsoonal rainfall occurs strongly in
February due to the climatological wind conditions that are normally cyclonic over
northwestern Australia.
1. Introduction
The sea surface temperature (SST) variability of the tropical Pacific has important implications for
climate worldwide, primarily associated with El Niño-Southern Oscillation (ENSO) events. The
eastern Pacific warm water associated with El Niño episodes induces an anomalous Walker cell that
leads to subsiding air and relative dry conditions over the maritime continent. When the maximum
SST warming is located in the central Pacific instead, an anomalous double Walker cell is observed in
the troposphere with the joint ascending branch over the warm anomaly and descending branches over
the eastern Pacific and the Indonesian region [1]. This “flavour“ of El Niño, termed El Niño Modoki,
has been associated with climate impacts distinct from conventional El Niño events (e.g., [1, 2, 3, 4]).
The Modoki phenomenon has become more frequent than traditional El Niños since the late 1970s [1].
c 2010 IOP Publishing Ltd
1
17th National Conference of the Australian Meteorological and Oceanographic Society
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 11 (2010) 012031
doi:10.1088/1755-1315/11/1/012031
Over Australia, dry conditions are observed in December and March during El Niño Modoki events
[5,6]. In addition, unusually wet conditions occur in northern Australia during the peak of the
monsoon season, particularly in February [6]. As the positive rainfall anomaly in February counteracts
the negative one in December, analysis of the averaged austral summer season therefore gives a false
impression that Modoki does not affect the Australian monsoon. This behaviour is the opposite to that
observed during traditional El Niño events and seems to be a robust feature across different datasets
(figure 1).
Over northwestern Australia, this bi-modal rainfall anomaly behaviour in austral summer indicates
that Modoki events tend to delay the normal monsoon onset and advance the monsoon retreat over the
region. Modoki does not necessarily reduce the total amount of rainfall received in the year, but it
leads to a rearrangement of the annual precipitation cycle in northwestern Australia, producing a
shorter and more intense monsoon season.
Jan
(d) BOM - ENSO
(c) BOM
(b) NCEP
(a) CMAP
Dec
DJFM
Mar
Feb
12oS
o
24 S
2
o
36 S
1.5
o
12 S
1
24oS
0.5
36oS
0
o
12 S
o
24 S
−0.5
o
36 S
−1
o
12 S
−1.5
o
24 S
−2
36oS
o
o
120 E 140 E
o
o
120 E 140 E
o
o
120 E 140 E
o
o
120 E 140 E
o
o
120 E 140 E
Figure 1. Composites of rainfall
anomalies during Modoki (a-c) and
conventional El Niño (d) events from
December to March and for the
December to March (DJFM) average
using the (a) CMAP, (b) NCEP/NCAR
Reanalysis and (c-d) BOM datasets.
Modoki years based on Ashok et al.
(2007): 1979/80, 1986/87, 1990/91,
1991/92, 1992/93, 1994/95, 2002/03,
2004/05. El Niño years: 1957/58,
1965/66, 1969/70, 1972/73, 1982/83,
1987/88, 1997/98. Units in mm day-1.
Areas within the thin black line are
statistically significant at the 90% level
according to a two-tailed t-test.
The reason why Australia experiences dry conditions with a warm anomaly in the equatorial
Pacific is already known to be due to Walker circulation changes and the anomalous subsiding air that
inhibits convection, the formation of clouds and thus reduces rainfall. An explanation for the opposite
rainfall response during the peak of the monsoon season has been proposed by [6]. Using observations
and numerical experiments, the authors have shown that northwestern Australian rainfall is enhanced
in February due to an increase in moisture convergence caused by an anomalous cyclonic circulation
over the region. However, the mechanisms responsible for generating the anomalous cyclonic
circulation were not established. More recently, [7] suggested that this circulation anomaly is a remote
response to anomalous heating in the tropical Pacific as developed in the Gill-Matsuno theory.
In this study we address the physical link between the warm Pacific SST anomalies associated with
Modoki events and the anomalous cyclonic circulation that causes enhanced rainfall in northwestern
Australia during February. This study complements [7] by considering the mean wind state and by
examining the initialization of the air-sea process in the AGCM that leads to a Gill-Matsuno response.
2. Data and methods
2.1. Data sets and climatologies
In this study we use the rainfall datasets from the Australian Bureau of Meteorology (BOM; [8]), the
CPC Merged Analysis of Precipitation (CMAP; [9]), and the National Center for Environmental
Prediction (NCEP) / National Center for Atmospheric Research (NCAR) Reanalysis [10]. In addition,
the global SST analysis from the Hadley Centre (HadISST1; [11]), the outgoing longwave radiation
2
17th National Conference of the Australian Meteorological and Oceanographic Society
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 11 (2010) 012031
doi:10.1088/1755-1315/11/1/012031
(OLR) from the NOAA Interpolated OLR [12] and horizontal winds from the NCEP/NCAR reanalysis
are also used here. The use of different datasets corroborates the analysis of [7] that is based on the
NCEP/NCAR Reanalysis only.
We focus our analysis on the period for which more reliable global data exists, and on the period
when Modoki events are more frequent, i.e. 1979 to 2006. The El Niño Modoki years used here are the
same as those selected by [1]: 1979/80, 1986/87, 1990/91, 1991/92, 1992/93, 1994/95, 2002/03 and
2004/05. The Modoki year starts in June and ends in May of the following year.
2.2. Numerical experiments
A series of four experiments generated with the NCAR Community Atmospheric Model (CAM3) was
performed by [6] to test the sensitivity of the Australian rainfall to the location of the SST warming
along the equatorial Pacific. Each experiment consisted of 50 members integrated for 12 months. A
1°C anomaly was superimposed onto the SST seasonal cycle around 10°S-10°N over four locations:
eastern, central-eastern, central-western and western Pacific. The authors concluded that Australian
rainfall is more sensitive to the warm SST anomaly in the central-western Pacific, suggesting that the
dateline is a key region for the Modoki forcing. Here we make use of this experiment, hereinafter
referred to as the central-western Pacific experiment.
3. Results
3.1. Seasonality of the atmospheric response: Observations
The fact that the northern Australian rainfall anomalies during Modoki events change sign during the
peak of the monsoon for the same tropical Pacific SST signature suggests that this response is phaselocked to the mean seasonal cycle. Figure 2(a-f) shows the long-term monthly mean rainfall, the
regions with strong convective activity (via low OLR values) and the location of the maximum SST in
the southern latitudes. The intensification of the precipitation rate and the broadening of the deep
convection in the southern latitudes are most pronounced in January and February, indicating the
strengthening of the South Pacific Convergence Zone (SPCZ).
To assess the seasonality of the atmospheric response during Modoki years, we show in figure 2(gi) the composite seasonal cycle for Modoki years and long-term mean seasonal climatology for the
southern tropical Pacific region around the dateline. This region is chosen as it contains the warm SST
anomalies in the key central-western Pacific region as suggested by [6], and also because it exhibits
strong convective and rainfall responses from the seasonally strengthened SPCZ. Although Modoki
events are associated with a statistically significant warming of the SST compared to the mean state,
the responses in precipitation and OLR are only significantly different from the climatology from
November to February. This demonstrates the seasonality in the interaction between the interannual
anomalies from Modoki events and the mean state of the atmosphere.
3.2. The Gill-Matsuno mechanism: Modelled response in February
We assess the simulated atmospheric response to the SST warming in the central-western Pacific for
February in figure 3. The positive SST anomaly changes the local surface flux balance via enhanced
evaporation. A zone of low pressure forms over the underlying warm waters, driving a convergence of
low-level winds over the central-western Pacific (figure 3a).
At the same time the anomalous SST forcing drives additional diabatic heating to the troposphere
via enhanced air-sea heat fluxes. In the equatorial regions, any available potential energy generated by
diabatic heating is immediately converted to kinetic energy. Therefore, rising motion is also
intensified by the convergence of low level winds (via continuity) and by enhanced diabatic heating
(via thermodynamics). The anomalous positive vertical velocity transports moisture from the
convective mixing layer to the mid troposphere, favouring the development of deep convective clouds.
Deep convection eventually leads to disturbances in the high troposphere generating a GillMatsuno-type response [13, 14] in the central-west Pacific and a remote response in the atmosphere
3
17th National Conference of the Australian Meteorological and Oceanographic Society
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 11 (2010) 012031
doi:10.1088/1755-1315/11/1/012031
via excitation of equatorial waves. The remote response to the SST forcing is seen over Australia with
a center of low pressure and anomalous cyclonic circulation (figure 3a). The cyclonic circulation
anomaly is also seen in the vertically integrated moisture flux from the surface to 500hPa (figure 3b),
indicating anomalous convergence of moisture over northern Australia. The anomalous convergence
of moisture and cyclonic circulation enhance the normal atmospheric conditions during the monsoon
season over Australia, i.e. a cyclonic circulation in February (figure 2m) and low pressure over the
continent, generating intensified convection and increased monsoonal rainfall over the northwest
region. Note that only during January and February is the mean circulation cyclonic over northwestern
Australia (figure 2l-m). The anomalous cyclonic circulation then reinforces the background state,
generating enhanced monsoonal rainfall during the peak of the Australian monsoon.
(a) Nov
(c) Jan
(b) Dec
15
o
20 N
o
0
o
20 S
10
o
120 E
o
160 E
o
160 W
o
120 E
o
160 E
o
120 E
o
160 W
o
160 E
(f) Apr
(e) Mar
(d) Feb
o
160 W
o
20 N
0
5
o
o
20 S
o
120 E
o
160 E
o
120 E
o
160 W
(g) SST - Celsius
o
160 E
o
120 E
o
160 W
(h) PREC - mm/day
12
240
8
29
28.5
Dec
6
220
4
Feb
Apr
Jun
Aug
Oct
(j) Nov
Dec
Feb
Apr
Jun
(k) Dec
Aug
Oct
200
Dec
Feb
Apr
0°
0°
0°
0°
10°S
10°S
10°S
20°S
20°S
20°S
20°S
140°E
30°S
180°
100°E
140°E
30°S
180°
100°E
140°E
Jun
Aug
Oct
(m) Feb
(l) Jan
10°S
30°S
100°E
0
o
160 W
(i) OLR - W/m2
260
10
29.5
o
160 E
180°
30°S
100°E
140°E
20 m/s
180°
Figure 2. (a-f) Long term monthly mean precipitation (shaded, mm day-1), OLR<220W/m2 (black
line) and maximum SST location (red line, °C) from November to April. (g-i) Annual cycles of
SST, precipitation (from CMAP) and OLR based on 1979-2006 mean (solid line) and Modoki
years (dashed-circle line) for the averaged southern tropical Pacific region around the dateline, i.e.
0°-10°S, 160°E-160°W. The gray areas represent the 95% confidence level based on a Monte Carlo
test. (j-m) Long term monthly mean winds at 850hPa from November to February.
The remote response, based on the Gill-Matsuno theory, is most clearly seen in the simulated
streamfunction anomaly fields (figure 3d-e). A horizontal quadrupole vortex structure, consistent with
[13], appears at the upper levels of the atmosphere with negative anomaly over Australia. An opposing
positive anomaly occurs at low levels over Australia, consistent with the anomalous cyclonic
circulation and low sea level pressure anomaly. This indicates a simulated baroclinic response in the
tropics during Modoki events. In addition to the return flow west of the heat forcing, the model
simulates an anomalous Walker circulation with upward motion in the central-western Pacific in
response to the warm equatorial SST (figure 3f). Mass conservation requires a compensating
4
17th National Conference of the Australian Meteorological and Oceanographic Society
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 11 (2010) 012031
doi:10.1088/1755-1315/11/1/012031
downward motion, which occurs over the Indian Ocean. Interestingly, Australia lies in between the
ascending and descending Walker cell anomalies (figure 3f), which allows the Gill return flow to
freely influence the rainfall over the continent.
o
25 N
(a) Sea Level Pressure and 850hPa Winds − Feb
1.5
1
0
o
0.5
o
25 S
(d) Streamfunction at 100hPa − Feb
60oN
6
x 10
6
4
30oN
2
0
0o
0
−0.5
o
−2
30 S
−4
−1
o
50 S
o
80 E
o
120 E
o
160 E
−1.5
o
120 W
o
160 W
(b) Moisture Flux and Divergence of Moisture − Feb
o
25 N
0
400
200
o
0
o
25 S
o
50 S
80oE
120oE
160 oE
160oW
120 oW
(c) Precipitation Anomaly − Feb
1
o
18 S
0
60 E
120 E
o
180 W
o
120 W
o
60 W
(e) Streamfunction at 850hPa − Feb
60oN
−6
6
x 10
3
2
30oN
1
o
0
0
o
−1
−200
30 S
−400
o
−2
60 S o
0
60oE
120oE
180oW
120oW
60oW
(f) Velocity Potential at 100hPa − Feb
−3
6
x 10
4
30 N
o
2
0o
0
−2
o
60 S o
0
−1
o
140 E
o
30oS
−0.5
o
120 E
o
o
60 N
0.5
o
36 S
60oS o
0
60oE
120oE
180oW
120oW
60oW
−4
Figure 3. Simulated anomalies in February for the idealized experiment forced with 1°C SST
warming in central-western equatorial Pacific (10°S-10°N, 160°E-160°W). (a) Sea level pressure
(hPa) and wind anomalies at 850hPa (m s-1). (b) Vertically integrated moisture flux (kg m-1s-1)
anomaly from the surface to 500hPa and the associated divergence (kg s-1). (c) Precipitation
anomaly (mm day-1). (d) Streamfunction anomaly at 100hPa (m2s-1). (e) Streamfunction anomaly at
850hPa (m2s-1). (f) Velocity potential anomaly at 100hPa (m2s-1). Adapted from [7].
(a) Day 2
80
o
18 N
(b) Day 3
o
80
18 N
o
o
0
0
0
0
o
18 S
18oS
o
100 E
o
o
150 E
o
160 W
(c) Day 4
100 E
80
o
18 N
o
o
150 E
o
160 W
(d) Monthly Mean
500
18oN
o
0
0
o
0
0
o
18 S
18 S
o
100 E
o
150 E
o
160 W
100oE
150oE
Figure 4. Evolution
of the anomalies of
sea level pressure
and winds at 850hPa
on (a) day 2, (b) day
3, (c) day 4, and (d)
over 30-60 days
average representing
the
February
anomaly.
160oW
A preliminary analysis of the daily response to the simulated SST warming in the equatorial Pacific
confirms the initialisation of the mechanism described here. Figure 4 shows the sea level pressure and
low-level wind anomalies for the first three days of the simulation, after the Modoki-like SST
5
17th National Conference of the Australian Meteorological and Oceanographic Society
IOP Publishing
IOP Conf. Series: Earth and Environmental Science 11 (2010) 012031
doi:10.1088/1755-1315/11/1/012031
perturbation is superimposed on the surface SST forcing on January 1st. The anomaly plotted in figure
4 is based on a single simulation and as such should be viewed with certain caution. However the
response for the first few days is not expected to vary considerably given the predictability of the
atmosphere to an initial SST forcing. The 30-60 days mean (representative of February) exhibits a
similar solution as that presented in figure 3a based on the 50-member ensemble monthly mean.
4. Conclusions
Northern Australia generally experiences dry conditions during both canonical El Niño and El Niño
Modoki events. This results from changes in the Walker circulation that drives anomalous subsidence,
weakened convection and reduced rainfall over the continent. However, during El Niño Modoki
events, when the location of the maximum SST anomaly is centered on the dateline, at the peak of the
monsoon season, an increase in rainfall over northwestern Australia generally occurs. The
teleconnection between the warm SST anomaly over the central-western Pacific and the intensified
monsoonal rainfall over northwestern Australia can be explained by the Gill-Matsuno theory and the
interaction between this interannual response and the climatological mean circulation state.
The short-lived rainfall increase only occurs in January-February phase-locked to the strengthening
of the South Pacific Convergence Zone (SPCZ). The intensification of the SPCZ in the southern
summer months is conducive to the development of increased deep convection and heavy rainfall in
the central-west Pacific. During Modoki years, the warmer than normal SST in the central-west Pacific
decreases the local pressure, causing anomalous wind convergence at low levels and a consequent
rising motion. In addition, the warm SST anomaly also enhances the supply of latent heat flux to the
moist mixing layer. This anomalous configuration during Modoki events, conspires to enhance the
convection that occurs during the normal SPCZ intensification during austral summer. Enhanced
vertical transport of moisture from the convective mixing layer to the mid troposphere drives intense
deep convection over the southern latitudes of the central-western tropical Pacific. Moist convection
and enhanced rainfall generates anomalous diabatic heating that eventually leads to disturbances in the
high troposphere akin to a Gill-Matsuno-type response. The Gill-Matsuno response is associated with
a remote cyclonic circulation over northwestern Australia. The anomalous circulation reinforces the
mean cyclonic winds over the region during January and February, thus enhancing the monsoon. As a
consequence, there is increased convergence of moisture and enhanced rainfall over the region.
References
[1] Ashok K, Behera S K, Rao S A, Weng H and Yamagata T 2007 J. Geophys. Res. 112
doi:10.1029/2006JC003798
[2] Kumar K K, Rajagopalan B, Hoerling M, Bates G and Cane M 2006 Science 314 115–9
[3] Wang G and Hendon H H 2007 J. Clim. 20 4211–26
[4] Weng H, Ashok K, Behera S K, Rao S A and Yamagata T 2007 Clim. Dyn. 29 113–29
[5] Taschetto A S and England M H 2009 J. Clim. 22 3167–74
[6] Taschetto A S, Ummenhofer C C, Sen Gupta A and England M H 2009 Geophys. Res. Lett. 36
L12704, doi:10.1029/2009GL038416
[7] Taschetto A S, Haarsma R D, Sen Gupta A, Ummenhofer C C, Hill K J and England M H 2010
J. Clim., in press
[8] Jones D and Weymouth G 1997 Bureau of Meteorology Technical Report 70 19 pp
[9] Xie P and Arkin P A 1997 Bull. Amer. Meteorol. Soc. 78 2539–58
[10] Kalnay E et al 1996 Bull. Am. Meteorol. Soc. 77 437–71
[11] Rayner N A, Parker D E, Horton E B, Folland C K, Alexander L V, Rowell D P, Kent E C and
Kaplan, A 2003 J. Geophys. Res. 108 doi:10.1029/2002JD002670
[12] Liebmann B and Smith C A 1996 Bull. Am. Meteorol. Soc. 77 1275–7
[13] Gill A E 1980 Q. J. R. Meteorol. Soc. 106 447–62
[14] Matsuno T 1966 J. Meteorol. Soc. Japan 44 25–43
6